77% of Diners Check Your Website Before They Visit. Can Google Read Your Menu?

Before a guest decides to visit or order, they check the menu. Not the Instagram, not the reviews, the menu. And 77% of diners visit a restaurant's website before they commit to eating there at all, according to a 2026 survey covered by Restaurant Dive. Here is the problem for most independent restaurants in Denver: the menu on that website is a photo, a PDF, or a page that only exists on a delivery app. To a search engine, and to the AI assistants a growing share of diners ask instead, that menu is close to blank.

What "Google Can't Read It" Actually Means

Google can index a PDF file. That is not the same as reading it. The text inside a PDF menu is not counted toward what your website is about, so the dish names, the descriptions, and the allergen notes do nothing to tell Google what your kitchen serves. An image menu is worse. Every word on it is locked inside a picture.

There is a test any owner can run in ten seconds. Search your restaurant's name plus a dish you are known for. If your own website does not come up, Google cannot see that you serve it.

The Assistants Answering "Where Should We Eat"

Google's AI Overviews mostly stay quiet on restaurant questions. The dining answers are coming from ChatGPT, Gemini, and Perplexity, which field "where should we eat tonight" and "what should I order there" constantly. Those tools quote menus that are written as text. A diner asking for a gluten-free entree under $20 nearby gets pointed at the restaurants whose menus can be read and quoted. A photo of a menu gives the assistant nothing to work with.

Heavy Menus Are Slow Menus

A menu built from three large photos is also a slow page on a phone. Restaurant sites already carry a lot: reservation widgets, ordering embeds, map frames, review scripts. Add uncompressed images on top and the page crawls. One 2026 analysis of restaurant site speed put fewer than half of all sites in the passing range for Google's Core Web Vitals, with website-builder and template sites the most likely to fail. A guest who taps through from Google and waits four seconds for a menu picture to load often taps straight back out.

The dishes already exist. The prices already exist. They just need to live on the page as words instead of inside a picture.

What a Readable Menu Looks Like

• Every dish as real text on the page, with a short description. Not "Green chili $9" but "Pork green chili with hominy, fire-roasted Hatch chiles, and a flour tortilla."
• One Menu block of structured data in JSON-LD, so Google and AI tools pull the sections, items, and prices directly.
• The same name, address, and phone number on the site, in the structured data, and on the Google Business Profile.
• The Google Business Profile menu field filled in, not left empty.
• A page a single dish can be linked to, so a review, a post, or an AI answer can point right at it.

One Audit, One Number

We recently audited an independent restaurant with a wall of five-star reviews and a room that fills on weekends. Good food, loyal regulars, a real business. Its menu page was most of a megabyte of website-builder code wrapped around three JPEG images. We searched that page for the dishes people drive across town to order. They appeared zero times. The prices appeared zero times. The Google Business Profile menu field was empty. None of that hurt the room, and all of it hurt the search.

The Fix Is Not a Project

Rebuilding a menu as text is a few hours of work. The dishes already exist, the prices already exist, and they just need to sit on the page as words. Do it once and the menu starts working in the places guests look first: the search result, the map listing, and the AI answer. Stale and unreadable menus are one of the seven items on our free visual bottlenecks checklist for that reason.

The menu is the page guests check first, and for most independent restaurants it is the page doing the least work. Fixing it pays back every time someone searches a dish, opens the map, or asks an assistant where to eat nearby.

Full local-search breakdown: restaurant SEO in Denver. The diner side of this shift: how restaurants get recommended by AI.

Questions Owners Actually Ask

Sources: Restaurant Dive, "77% of diners check restaurant websites before visiting"; The Forking Group, "Restaurant Menu SEO"; The Hospitality Hangout, "AI Restaurant Discovery" (2026); RichMenu, "Restaurant Website Speed in 2026".